ORIGINAL UNEDITED MANUSCRIPT

Transcription

1 1 Combining a Food Frequency Questionnaire with 24-Hour Recalls to Increase the Precision of Estimating Usual Dietary Intakes Evidence from the Validation Studies Pooling Project Laurence S. Freedman, Douglas Midthune, Lenore Arab, Ross L. Prentice, Amy F. Subar, Walter Willett, Marian L. Neuhouser, Lesley F. Tinker, and Victor Kipnis Correspondence to: Dr. Laurence Freedman, Biostatistics Unit, Gertner Institute for Epidemiology, Sheba Medical Center, Tel Hashomer 52621, Israel ( lsf@actcom.co.il) ; telephone number: ; fax number: Author affiliations: Biostatistics Unit, Gertner Institute for Epidemiology, Sheba Medical Center, Tel Hashomer, Israel (Laurence S. Freedman); Information Management Services, Inc., Rockville MD (Laurence S. Freedman); Biometry Research Group, Division of Cancer Prevention, US National Cancer Institute, Bethesda, MD (Douglas Midthune, Victor Kipnis); Division of General Internal Medicine, David Geffen School of Medicine, University of California Los Angeles, Los Angeles, CA (Lenore Arab); Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA (Ross L. Prentice, Marian L. Neuhouser, Lesley F. Tinker); Division of Cancer Control and Population Sciences, US National Cancer Institute, Bethesda, MD (Amy F. Subar); Departments of Epidemiology and Nutrition, Harvard School of Public Health, Boston, MA (Walter Willett); Channing Division of Network Medicine, Department of Medicine, Brigham and Women s Hospital, Boston, MA (Walter Willett); Harvard Medical School, Harvard University, Boston, MA (Walter Willett). Funding: This work was supported by the following sponsors: The National Heart, Lung, and Blood Institute at the National Institutes of Health, U.S. Department of Health and Human Services [contracts HHSN C, HHSN C, HHSN C, HHSN C, HHSN C, and HHSN C to the Women s Health Initiative program]; the National Cancer Institute [grant R01 CA to the Nutrition and Physical Activity Assessment Study project]. Conflict of interest: Lenore Arab has IP interests in DietDay, the web-based 24-hour recall instrument used in the Energetics study. All of the other authors declare no conflict of interest. Running Head: Combining FFQs with 24-Hour Recalls Abbreviations: AMPM Automated Multiple Pass Method (Study); FFQ food frequency questionnaire; NBS Nutrition Biomarker Study; NPAAS Nutrition and Physical Activity Assessment Study; OPEN Observing Protein and Energy (Study); WHI Women s Health Initiative; 24HR Twenty-four hour recall. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health This work is written by (a) US Government employee(s) and is in the public domain in the US.

2 2 Abstract: Improving estimates of individuals dietary intakes is key to obtaining more reliable evidence for diet-health relationships from nutritional cohort studies. One approach to improvement is combining information from different self-report instruments. Previous work evaluated the gains obtained from combining information from a food frequency questionnaire (FFQ) and multiple 24-hour recalls (24HRs), based on assuming that 24HRs provide unbiased measures of individual intakes. Here, we evaluate the same approach of combining instruments, but based on the better assumption that recovery biomarkers provide unbiased measures of individual intakes. Our analysis uses data from the five large validation studies included in the Validation Studies Pooling Project: the Observing Protein and Energy study ( ), Automated Multiple Pass Method validation study (2002-4), Energetics study (2006-9), Nutrition Biomarker Study (2004-5) and Nutrition and Physical Activity Assessment Study (2007-9). The data include intakes of energy, protein, potassium, and sodium. Under a time-varying usual intake model analysis, combining a FFQ with 4 24HRs improved correlations with true intake for predicted intakes of protein density, potassium density and sodium density (range ), over a single FFQ (range ). Absolute increases in correlation ranged from 0.02 to 0.26, depending on nutrient and sex, with an average increase of Based on unbiased recovery biomarker evaluation for these nutrients, we confirm that combining a FFQ with multiple 24HRs modestly improves the accuracy of estimates of individual intakes. Keywords: Cohort studies, Dietary measurement, Energy, Measurement error, Potassium, Protein, Recovery biomarkers, Sodium

3 3 Measurement error in self-reported dietary intakes limits the reliability of results from nutritional cohort studies 1. Approaches to improving the accuracy of dietary intake estimates include reporting intake using new technologies 2-3, combining self-reports with biomarkers 4 and combining different self-report instruments 5. We focus on the last. Carroll et al 5 considered combining information from multiple 24-hour recalls (24HRs) and a food frequency questionnaire (FFQ), using data from the Eating at America s Table Study 6. Their evaluation assumed that 24HRs unbiasedly measure an individual s intake, and they showed their results were probably robust to departures from this assumption. Here, we also evaluate combining 24HRs with a FFQ, but under the better-founded assumption that recovery biomarkers unbiasedly measure individual intake 7. We use data from five large validation studies with recovery biomarkers as reference instruments, comprising the Validation Studies Pooling Project 8. Also motivating this work was previous work using the time-varying intake model 9, which showed that correlations between intakes reported using 24HRs and true usual intake were lower than previously estimated, due to the close proximity in time of biomarker and 24HR assessments in some studies 10. This raised doubts regarding the benefit accruing from adding 24HR assessments to a FFQ. Using recovery biomarker and self-report data, we have estimated correlations of intakes estimated from self-report instruments and their combinations with true longer-term usual intakes for energy, protein, potassium, sodium and their densities. We report that combining instruments led

4 4 to increased correlation of estimated with true intakes and discuss implications of these findings. Determining the optimal combination of 24HR and FFQ data requires knowing the error models of the two instruments, which in turn requires recovery biomarker measurements. Here, we focus on cohort studies having FFQ and 24HR but no biomarker data. The FFQ and 24HR data would then be combined, as the best option available, under the (usually erroneous) assumption that 24HRs are unbiased. We use recovery biomarkers to provide an unbiased assessment of this method of combining instruments. MATERIALS AND METHODS The Validation Studies Pooling Project Investigators of five large (>200 participants) validation studies using recovery biomarkers agreed to pool their data, aiming, through common analysis, to clarify the nature and magnitude of reporting errors in FFQs and 24HRs. 8,11 The five Validation Studies Pooling Project studies included diverse populations within the US. The Observing Protein and Energy (OPEN) study 12 and the Automated Multiple Pass Method (AMPM) validation study 13 included volunteers, aged 40-69y (OPEN) or 30-69y (AMPM), residing in Maryland. The Energetics study included younger white and African-American adults residing in California. 14 The Nutrition Biomarker Study (NBS) 15 and the Nutrition and Physical Activity Assessment Study (NPAAS) 16 included postmenopausal women, mostly over 60y and residing throughout the US, in the

5 5 Women s Health Initiative (WHI) Dietary Modification Trial and the WHI Observational Cohort, respectively. Further details are in references At least one FFQ was administered to each participant. This analysis includes only the first administration. The FFQs queried intake over the past year (OPEN, Energetics, AMPM) or three months (NBS, NPAAS). One of three versions was used, the Harvard FFQ 17 (AMPM), the WHI FFQ 18,19 (NBS, NPAAS) or the Diet History Questionnaire 6 (OPEN, Energetics). Each study included two or more 24HR assessments, administered to all participants in four studies, and to a 20% subset in NBS. An interviewer-administered multiple-pass method was used (4 studies), or a web-based self-administered 24HR (Energetics). Further details are in references Each study measured doubly-labeled water for energy intake; 20 and collected 24-hour urine for measuring nitrogen, 21 potassium 22 and sodium 23 intakes. Details of the laboratories and methods are in Web Appendix 1. Urinary nitrogen in grams was divided by 0.81 to convert to dietary nitrogen, 21 and then multiplied by 6.25 to convert to dietary protein. Urinary potassium was divided by 0.8 to convert to dietary potassium, 24 and urinary sodium by 0.86 to convert to dietary sodium. 25

6 6 The timing of measurements varied across studies (Web Figure 1). In four studies, the FFQ was administered at entry, but in AMPM, 1-14 months after entry. In all studies, doubly-labeled water was administered at 1-14 days, and 24-hour urines collected during the same period. The 24HRs were administered on the following days: OPEN 1 and 61; Energetics 1, 3, 5, 8, 11, 14, 30 and 60; AMPM 1, 5-6, 10-11; NBS twice at around 180 days in a 20% sub-sample; NPAAS three administrations from 14 to 104 days. These measurements were supplemented by sub-studies, of varying size, to examine reliability of self-reports and biomarkers. The time between main and sub-study administrations ranged from 2 weeks in OPEN, to approximately 6 months in Energetics, NBS and NPAAS, and months in AMPM. In OPEN only doubly-labeled water was repeated; in other studies all biomarkers and self-reports were repeated. For example, in NBS and NPAAS, the entire study protocol was repeated in a 20% subsample. Our analysis included repeat biomarker and 24HR assessments. Statistical methods We report on seven dietary components: energy; protein, potassium, and sodium, and their densities (ratios to energy intake). We predicted targeted true usual intake (defined here as the 12-month average) using different self-report instruments. A measure of the goodness of a prediction is its correlation with truth. The correlation measures how well the prediction orders

7 7 individuals according to their true intake and relates to loss of statistical power. 26 Low values, e.g., less than 0.4, are undesirable, although there is no sharp cut-off. The time-varying intake model accounts for serial correlation in individuals intakes and proximity of self-report to biomarker assessments. 9 We describe briefly the model below. For technical details, see Web Appendix 2 and reference 9. Each sex and dietary component was modeled separately. Dietary variables were logarithmically transformed. The model comprised four meta-analysis sub-models with study-specific parameters. The first three sub-models specified linear regression relationships between the biomarker, 24HR and FFQ, respectively, and true intake, the explanatory variable. The fourth sub-model specified how true intake varied over time. The time axis was divided into 90-day sub-periods. Biomarker sub-model: Biomarkers were assumed to measure true intake on the previous day (or for energy during the day assessment) without bias (intercept=0, slope=1) but with independent random error. The error variance was estimated through repeat assessments performed within the same sub-period, and was assumed equal across studies, because of insufficient replications to provide study-specific estimates. 24HR sub-model: 24HR-reported intake was assumed to measure true intake on the same day with systematic intake-related bias. Person-specific bias and within-person random

8 8 error terms were included, as in Kipnis et al. 27 Within-person random error terms were assumed mutually independent, and independent of all other terms in the model. FFQ sub-model: FFQ-reported intake was assumed to measure true average intake over the past year with systematic intake-related bias. With no FFQ replications, personspecific bias and within-person random error terms could not be separately estimated, and were combined as a total error term. This total error was assumed correlated with the 24HR person-specific bias. Time-varying true intake sub-model: We assumed that: (i) an individual s true intake varied over time, but the group average and variance, on a single day and in each sub-period, remained constant within study; (ii) the ratio of the single-day variance to the 90-day usual intake variance (the intakevariance ratio) was common across studies; (iii) the correlation structure between usual intakes in different sub-periods was common across studies, and was autoregressive of order 1 (AR(1)) or compound symmetry, decided according to the best model-fit, using Akaike s Information Criterion. Model parameters were estimated using maximum likelihood, assuming log biomarker, 24HR and FFQ values to be normally distributed, using the Statistical Analysis System (SAS) 28 CALIS procedure.

9 9 Instruments were combined through a prediction (or calibration) equation for dietary intake, using 24HRs as reference. Via estimated model parameters, we estimated correlations of these predictions with truth, separately for each study, sex and dietary component, for 24HRs, FFQ or a combination. We present results for a single FFQ, and 1, 2, 4 or 8 24HRs with or without a FFQ. The estimates depend on the timing of selfreports. The timing assumed is shown in Web Table 1, and the calculation methods are in Web Appendix 2. Study-specific estimates are presented in the Web materials (specified below), and across-study summaries (weighted averages) in the main paper. RESULTS Details of the models fitted and parameter estimates are in Web Tables Table 1 shows for each dietary component, the estimated correlations with truth, averaged across studies, for a single FFQ, different numbers of 24HR administrations, and their combination. Correlations increased with the number of 24HR administrations, but with diminishing benefits. Similarly, adding a FFQ to one or more 24HRs improved correlations, but benefits diminished with increasing number of 24HRs. Benefits of combining 24HRs with a FFQ were seen both for absolute intakes and densities, but appeared smaller for women than men. Estimated correlations remained low (<0.40) for energy among men and women, and for sodium and sodium density among women. For other dietary components, correlations above 0.5 were attained for some combinations, but for protein density in women the maximum was Study-specific correlations are presented in Web Tables

10 10 We considered the least combination (the combination with the smallest number of self-report administrations) yielding a correlation at least 90% of that achieved by using 1 FFQ plus 8 24HRs (Table 2). One FFQ plus 3 or 4 24HRs provided the majority of the benefit from combining instruments. DISCUSSION Adding repeat administrations of a 24HR to a FFQ increased the accuracy of prediction of dietary intake, but with diminishing returns as the number of 24HR administrations increased. One FFQ plus 4 24HR administrations gave at least 90% of the maximum correlation. Using this combination, correlations between predicted and true intakes were for nutrient densities compared to for a single FFQ. Absolute increases in correlation ranged from 0.02 to 0.26, depending on nutrient and sex, with an average increase of These results strengthen those of Carroll et al 5, being based on unbiased recovery biomarker evaluation. Improvements in correlation from adding a single FFQ to four administrations of a 24HR were mostly marginal. However, the dietary components we studied are not episodically consumed. Carroll et al 5 also reported only small increases in correlations for such components.

11 11 The correlations obtained by combining 24HRs with a FFQ leave room for considerable improvement. Clearly, continued research into ways of assessing dietary intake is needed, so as to increase correlations to above 0.7. Neuhouser et al 15 showed that for energy and absolute protein intake, correlations can be considerably increased by including in the prediction equation body mass index (BMI) and other personal characteristics that are sources of systematic reporting error. More extensive data analysis confirms this and extends the result to absolute sodium intake and potassium density, with modest gains for potassium, protein density and sodium density 8,11. Such prediction equations hold promise of further gains in precision of measuring intakes. However, such prediction is effective only when derived from biomarker references. Here, we have focused on studies collecting a FFQ and 24HRs, where prediction equations are based on 24HRs as reference instruments. Such equations do not capture the importance of personal characteristics such as BMI 29. Consequently, prediction equations based on 24HRs as reference do not greatly improve correlations with truth (Web Tables 21-22). Biomarker-based prediction equations are preferable, but require cohort studies to include validation sub-studies with such biomarkers, and are limited to the few nutrients with suitable biomarkers.

12 12 A central assumption behind our modeling is that recovery biomarkers are unbiased for individual intake, and that their errors are random. In Web Appendix 2 we discuss this assumption. The results presented here are for the limited set of dietary components that have recovery biomarkers. Carroll et al 5 have shown that combining 24HRs and a FFQ can lead to improvements across a wider range of dietary components. Together, this suggests that while the gains we have demonstrated will not apply to all dietary components, they will apply to many nutrients and foods. Acknowledgments We thank the following people for their valuable contributions to this work: Dr. Alfonso Ang (University of California, Los Angeles), John Commins (Information Management Services), Dr. Sheila Bingham (deceased, Dunn Nutrition Unit, British Medical Research Council, Cambridge), Dr. Kevin Dodd (US National Cancer Institute), Dr. Alanna Moshfegh (United States Department of Agriculture), Dr. Nancy Potischman (National Institutes of Health), Dr. Dale Schoeller (University of Wisconsin), and Dr. Richard Troiano (National Institutes of Health). We also thank the Women s Health Initiative investigators and staff for their dedication. A full listing of Women s Health Initiative investigators can be found at tigator%20short%20list.pdf.

13 13 REFERENCES 1. Subar AF, Freedman LS, Tooze JA, et al. Addressing current criticism regarding the value of self-report dietary data. J Nutr. 2015; 145 (12): Boushey CJ, Spoden M, Zhu FM, et al. New mobile methods for dietary assessment: review of image-assisted and image-based dietary assessment methods. Proc Nutr Soc. 2017; 76 (3): Stumbo PJ. New technology in dietary assessment: a review of digital methods in improving food record accuracy. Proc Nutr Soc. 2013; 72 (1): Freedman LS, Midthune D, Carroll RJ, et al. Using regression calibration equations that combine self-reported intake and biomarker measures to obtain unbiased estimates and more powerful tests of dietary associations. Am J Epidemiol. 2011; 174 (11): Carroll RJ, Midthune D, Subar AF, et al. Taking advantage of the strengths of two different dietary assessment instruments to improve intake estimates for nutritional epidemiology. Am J Epidemiol. 2012; 175 (4): Subar AF, Thompson FE, Kipnis V, et al. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires: the Eating at America s Table Study. Am J Epidemiol. 2001; 154 (12): Kaaks R. Biochemical markers as additional measurements in studies of the accuracy of dietary questionnaire measurements: conceptual issues. Am J Clin Nutr. 1997; 65 (suppl): 1232S 1239S.

14 14 8. Freedman LS, Commins JM, Moler JE, et al. Pooled results from five validation studies of dietary self-report instruments using recovery biomarkers for energy and protein intake. Am J Epidemiol. 2014; 180 (2): Freedman LS, Midthune D, Dodd KW, et al. A statistical model for measurement error that incorporates variation over time in the target measure, with application to nutritional epidemiology. Stat Med. 2015; 34 (27): Freedman LS, Midthune D, Carroll RJ, et al. Application of a new statistical model for measurement error to the evaluation of dietary self-report instruments. Epidemiology 2015; 26 (6): Freedman LS, Commins JM, Moler JE, et al. Pooled results from five validation studies of dietary self-report instruments using recovery biomarkers for potassium and sodium intake. Am J Epidemiol. 2015; 181 (7): Subar AF, Kipnis V, Troiano RP, et al. Using intake biomarkers to evaluate the extent of dietary misreporting in a large sample of adults: the OPEN study. Am J Epidemiol. 2003; 158 (1): Moshfegh AJ, Rhodes DG, Baer DJ, et al. The US Department of Agriculture Automated Multiple-Pass Method reduces bias in the collection of energy intakes. Am J Clin Nutr. 2008; 88 (2): Arab L, Wesseling-Perry K, Jardack P, et al. Eight self-administered 24-hour dietary recalls using the internet are feasible in African Americans and Caucasians: The Energetics Study. J Am Diet Assoc. 2010; 110 (6):

15 Neuhouser ML, Tinker L, Shaw PA, et al. Use of recovery biomarkers to calibrate nutrient consumption self-reports in the Women's Health Initiative. Am J Epidemiol. 2008; 167 (10): Prentice RL, Mossavar-Rahmani Y, Huang Y, et al. Evaluation and comparison of food records, recalls and frequencies for energy and protein assessment by using recovery biomarkers. Am J Epidemiol. 2011; 174 (5): Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol. 1985; 122 (1): Patterson RE, Kristal AR, Tinker LF, et al. Measurement characteristics of the Women s Health Initiative Food Frequency Questionnaire. Ann Epidemiol. 1999; 9 (3): Women's Health Initiative: Food Questionnaire. oc=/researchers/studydoc/whi%20forms/f060%20v1.6.pdf&action=default. Published Last updated May Accessed August Schoeller DA, Hnilicka JM. Reliability of the doubly labeled water method for the measurement of total daily energy expenditure in free-living subjects. J Nutr. 1996; 126 (1): 348S-354S. 21. Bingham SA, Cummings JH. Urine nitrogen as an independent validatory measure of dietary intake: a study of nitrogen balance in individuals consuming their normal diet. Am J Clin Nutr. 1985; 42 (6):

16 Mickelsen O, Makdani D, Gill JL, et al. Sodium and potassium intakes and excretions of normal men consuming sodium chloride or a 1:1 mixture of sodium and potassium chlorides. Am J Clin Nutr. 1977; 30 (12): Luft FC, Fineberg NS, Sloan RS. Estimating dietary sodium intake in individuals receiving a randomly fluctuating intake. Hypertension 1982; 4 (6): Freedman LS, Midthune D, Carroll RJ, et al. Adjustments to improve the estimation of usual dietary intake distributions in the population. J Nutr. 2004; 134 (7): Holbrook JT, Patterson KY, Bodner JE et al. Sodium and potassium intake and balance in adults consuming self-selected diets. Am J Clin Nutr. 1984; 40 (4): Kaaks R, Riboli E, van Staveren W. Calibration of dietary intake measurements in prospective cohort studies. Am J Epidemiol. 1995; 142 (6): Kipnis V, Subar AF, Midthune D, et al. The structure of dietary measurement error: results of the OPEN biomarker study. Am J Epidemiol. 2003; 158 (1): SAS Institute Inc.: Statistical Analysis System (SAS) software, Version 9.2. SAS Institute Inc. Cary, NC, US; Freedman LS, Commins JM, Willett W, et al. Evaluation of the 24-hour recall as a reference instrument for calibrating other self-report instruments in nutritional cohort studies: evidence from the Validation Studies Pooling Project. Am J Epidemiol. 2017; 186 (1):

17 Table 1: Correlation coefficients with truth of intakes predicted a from various combinations of self-report instruments b, combined over five validation studies c, for selected dietary components. Dietary Component No. 24HRs Sex FFQ (Y/N) Energy Male Y N Female Y N Protein Male Y N Female Y N Potassium Male Y N Female Y N Sodium Male Y N Female Y N Protein Density. Male Y N Female Y N Potassium Density Male Y N Female Y N Sodium Density Male Y N Female Y N

18 Abbreviations: 24HR: 24-hour recall; AMPM: Automated Multiple Pass Method; FFQ: Food frequency questionnaire; NBS: Nutrition Biomarker Study; NPAAS: Nutrition and Physical Activity Assessment Study; OPEN: Observing Protein and Energy a formed on the assumption that 24HRs are unbiased measures of intake b FFQ, 24HRs (1,2,4 or 8), or a combination of these c OPEN ( ), Energetics (2006-9), AMPM (2002-4), NBS (2004-5), NPAAS (2007-9) 18

19 19 Table 2: Least Combination (FFQ + 24HRs) that achieves correlation of at least 90% of the correlation for 1 FFQ HRs: OPEN ( ), Energetics (2006-9), AMPM (2002-4), NBS (2004-5), NPAAS (2007-9) Dietary Component Men Women Energy Protein Potassium Sodium Protein Density Potassium Density Sodium Density Abbreviations: 24HR: 24-hour recall; AMPM: Automated Multiple Pass Method; FFQ: Food frequency questionnaire; NBS: Nutrition Biomarker Study; NPAAS: Nutrition and Physical Activity Assessment Study; OPEN: Observing Protein and Energy